> For the complete documentation index, see [llms.txt](https://docs.degirum.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.degirum.com/orca/readme.md). # Overview This page provides an overview of the DeGirum Orca AI accelerator, describing its performance characteristics, support for pruned models, dedicated DRAM feature, and its flexible architecture. DeGirum® Orca is a flexible, efficient, and cost-effective AI accelerator. It helps developers build feature-rich edge solutions while staying within power and cost constraints. ## High Performance Orca's efficient architecture delivers strong real-world performance. A single Orca can handle multiple input streams and several ML models. See our [Orca Performance Benchmarks](/orca/benchmarks.md) for performance details. ## Support for Pruned Models Processing pruned models effectively boosts compute and bandwidth resources, letting you run larger, more accurate models in real time at the edge. ## Dedicated DRAM Dedicated DRAM helps applications quickly switch between ML models without lengthy transfers from the host. This reduces model-switching delays and is especially helpful when your application needs to change models often, such as in image or speech recognition scenarios. ## Flexible Architecture Orca's flexible architecture supports both int8 and float32 precision, so you can choose the format that best fits your use case and optimize performance, accuracy, and power consumption. {% embed url="" %}